can outliers be detected if use likert scale?

maeu

New Member
#1
Dear moderator,

My research comprises of 20 items of indeped. variables (e-com implementation barriers ) and 10 items of dep variables (involvement level in procurement activities using e-com) . Ind variables were tested using likert scale (1= strongly disagree to 5= strongly agree) and Dep Variables were tested using 1-5 scales (1= never involve to 5= very often involve).

Later, i plan to do Factor analysis, correlation and regression .

Based on the nature of my ind and depend. variables , here is my questions:

So far , i read many books and notes from the Internet and i found most outliers could be detected if the data are continous variables (like sales amount, IQ level, durations) . I haven't find any book shows how to detect outliers if the data are likert scale. Can outliers be detected if one uses likert scale?

If yes, could you teach me how to do detect outliers step by step using SPSS? Really appreciate prompt respond here or through my email : me_ribena_berry@yahoo.com .

Thank you ..

:wave:
MAEU

 
E

elnaz

Guest
#2
Hello
i am not moderator , but in talkstats every person help another
about your question , it is to be noted that if you do regression and correlation related to kind of data is ordinal or nominal or continous variable ( nor of ordinal and nominal )
i think regression is true only for continous variable and since your data are ordinal then for checking their relation you do
appropriate correlation ( Spearman ) between variables and use nonparametric tests.
i hope you will be succeed.
 

maeu

New Member
#3
Hi Elnaz,

Thank you for your concern and responding to my questions.
I met some statisticians in my uni , and they recommended that likert scale data can be used for parametric test like regression , correlation etc if i converted the value to Z score ( save standardised z score value as variables in descriptive functions)

Once all variables in likert scale have been converted to z score , i can proceed with parametric analysis..

The next step, i am not really sure whether i should check for outliers or not since i am using likert scale data. can or cannnot ?

If i still need to detect for outliers , i should use original data of likert scale or i need to use z score data of the likert scale? :confused:
 
E

elnaz

Guest
#4
Hello
it's wonderful , i will ask of my professor ,then i will tell you because of i dont agree with anything that you said me .

thanks a lot